|
Dear all,
I am working on an autologistic spatial lattice model (binary response data, neighbor matrix, and one explanatory variable) and would need help in the following aspects: - I would appreciate any references on autologistic spatial lattice models. - I used maximum pseudo-log-likelihood estimation and Monte Carlo simulation of parameters’ distributions to obtain parameter estimates. I would like to know if it is possible to estimate some sort of confidence interval around each of the parameters as well as perhaps some sort of testing statistics for significance of the parameters? (I know how to do that for SAR and CAR models but not for autologistic.) - Finally, I know that “large” and positive estimate of parameter on neighbor effects means that neighbors are statistically similar, but I was wondering if somebody could give me a range (based on their own experience and/or literature) of what “large” would be in this case (close to one perhaps?)? I would appreciate any help. Thank you very much in advance. Sincerely, Branka Turcin ------------------------------------------
Branka Turcin, M.Sc. Research Assistant Frontier Research System for Global Change International Arctic Research Center University of Alaska Fairbanks Phone: (907) 474 5730 Email: [EMAIL PROTECTED] http://www.frontier.iarc.uaf.edu:8080/~bturcin/ ----------------------------------------------- |
